[Rarius]

Can I ask you all to sanity check my design.

Requirements
I want a CNC router to be able to engrave PCBs, cut wood, plastics and aluminium.

History
I spotted a set of part on eBay, including 6 rails 4 spindles and all the linear bearings. I checked against the manufacturers website and found it was almost £200 worth. I got it for £55 posted! RESULT!

X-axis
805mm long 20mm round rails (http://www.worldofcnc.com/products.asp?recnumber=14), 660mm apart, running through 20mm linear bearings (http://www.worldofcnc.com/products.asp?recnumber=16). Driven by a 12mm trapezoidal spindle (http://www.worldofcnc.com/products.asp?recnumber=61). The spindle is stationary and the nut is driven.
To stop racking along this axis, I will add cables such as are used on draghtsmands drawing boards. I am sure this system has been discussed elsewhere on this site.

Y-axis
610mm long 16mm round rails (http://www.worldofcnc.com/products.asp?recnumber=14), 130mm apart, running through 16mm linear bearings (http://www.worldofcnc.com/products.asp?recnumber=16). Driven by a 12mm trapezoidal spindle (http://www.worldofcnc.com/products.asp?recnumber=61). The spindle is stationary and the nut is driven.

Z-axis
455mm long 12mm round rails (http://www.worldofcnc.com/products.asp?recnumber=14), 120mm apart, running through 12mm linear bearings (http://www.worldofcnc.com/products.asp?recnumber=16). Driven by a pair of 12mm trapezoidal spindles (http://www.worldofcnc.com/products.asp?recnumber=61) The spindles are attached to the Z carriage and driven by rotating nuts.

Notes
All motors are Nema23 size.
The router sits in the centre of the Z carriage.
I intend to build it all out of aluminium, but some parts may be made from wood/MDF to start with then replaced with aluminium parts made with the router.

[tightredcorset]

I have seen a lot of designs from home made. The biggest hurdle from any design will be the amount of flex in each axis vs feed speed. As torque is applied the rails must resist flex.

Your design looks sound. The proof of concept will be when you start that first cut. Aluminum will be the hardest. Keep the feed speeds low and I don't see any problems.

Let us know how it turns out.

Lee

[dan_the_welder]

Hi,

That looks cool. I would suggest a stiffener for the gantry.

Now the lead screw is serving that function. what I mean is when the gantry rails try to flex, they can pivot around the x rail, the only thing stopping this is the y lead screw. Which will be acting on your y lead screw bearings.

If you uses a piece of stock to join the two x carriages then the y rails may still flex, but not as much.

Also since your design is driving only one side of the x axis this stiffness will help keep the other side from binding or lagging.

Not a big deal for PCBs and light stuff but important for aluminum and harder materials.

Have Fun!

Dan

[Rarius]

Hi,

That looks cool. I would suggest a stiffener for the gantry.

Now the lead screw is serving that function. what I mean is when the gantry rails try to flex, they can pivot around the x rail, the only thing stopping this is the y lead screw. Which will be acting on your y lead screw bearings.

The Y lead screw doesn't move. All three leadscrews are fixed while the nuts rotate and draw themselves along the leadscrews.

I appreciate that I may need to strengthen the gantry. I'll look at doing that.

Also since your design is driving only one side of the x axis this stiffness will help keep the other side from binding or lagging.

As I explained in the text, although I am only driving one side of the X-axis, the other side is moved by a cable system similar to that used on a draughtsmans drawing board, so racking shouldn't be a serious problem.

[dan_the_welder]

Hi Rarius,

Oh Duh! No end bearings! *Face palm*

So just another plain rod in the missing corner might be all you need, I missed the cable anti racking thing too.

I like rotating nuts, I have never tried it myself and was curious to see that executed.

Dan

[Rarius]

Hi Rarius,
Oh Duh! No end bearings! *Face palm*

So just another plain rod in the missing corner might be all you need, I missed the cable anti racking thing too.

Easily done. I'm just thankful to have someone else casting an eye across this design.

I like rotating nuts, I have never tried it myself and was curious to see that executed.
Dan

I just finished making the first rotating nut. The original nuts are 24mm long and 26mm diameter. I have turned it down to 20mm and fitted a bearing and 30 tooth pulley onto it. Looks OK.. now I only need to make three more!

Here is an image of the design with the added stiffening on the gantry.

[MrRobot]

Looks good to me.

[FandZ]

I really like the belt driven nuts. I wanted try that on my build but decided against it for my first build. For a first or even 3rd design i think you did great. That said, I think you need to work on it some more.


I'm not sure if I'm fully grasping your design, but the way I'm seeing it, your spindle will be dropping down from the middle correct? I'm not seeing it upside down? If this is right side up, I'm curious why you have so much room above your z axis?

It looks to me you'll only have a z travel of maybe 2 inches. That would limit your choices of tooling bits a good bit. No pun intended :P

I see you are considering driving the Y axis from one side. Even with a cable support system, you'll be robbing your steppers of power. It can work, but why start with an inferior way of driving it? Steppers are cheap. My suggestion would be to either drive it from the middle (not possible with this current design) or drive it from both sides (the best option for all builds). You could also drive it from the middle useing belts but why waste stepper/servo power...

Lastly, it's a closed box. Which will limit the length of your stock. Consider opening the front and the back.

My last question would be, how do you plan to make this and out of what? I'm not talking about the rails and such but am really asking about the framing and supports. Will you be wielding it together or bolting?

[Rarius]

To make explanations easier I have rerendered the design with the bed and I have colour coded various sections of the design.

Frame = Grey
X Carriage (Gantry) = Red
Y Carriage = Yellow
Z Carriage = Blue
Spindle (Router) = Green

I'm not sure if I'm fully grasping your design, but the way I'm seeing it, your spindle will be dropping down from the middle correct? I'm not seeing it upside down? If this is right side up, I'm curious why you have so much room above your z axis?

You are not seeing it upside down. The image shows the machine from above.

I think I see where the confusion is coming from. What I had drawn was just the main workings of the router. As can be seen from the new image, it will all be mounted 12-18in above the bed. The spindle (green) will be mounted on the bottom plate of the Z axis (blue) and the body extends up through a hole in the center of the Y carriage (yellow). The bit will extend down through the hole in the bottom of the Z carriage.

It looks to me you'll only have a z travel of maybe 2 inches. That would limit your choices of tooling bits a good bit. No pun intended :P

I have a full 12in of Z travel. As the router is mounted on the bottom plate and moves through the center of the Y carriage, I can move the Z carriage all the way up and down. Of course I don't need this much travel most of the time, so will work much closer to the Y carriage.

I see you are considering driving the Y axis from one side. Even with a cable support system, you'll be robbing your steppers of power. It can work, but why start with an inferior way of driving it? Steppers are cheap. My suggestion would be to either drive it from the middle (not possible with this current design) or drive it from both sides (the best option for all builds). You could also drive it from the middle useing belts but why waste stepper/servo power...

Originally I was going to do just that, but after reading this forum for a month or two I rethought the system. From what I read, it seems that two motors on the same axis is just asking for trouble. A missed step on one motor will introduce forced racking.

As far as I have read (both here and elsewhere), the cable system is simple and will transfer power from one end of the gantry to the other with very little loss.

Lastly, it's a closed box. Which will limit the length of your stock. Consider opening the front and the back.

The frame as was drawn will be 400mm above the bed so there is plenty of room to move materials in and out. This can be seen in the new image.

My last question would be, how do you plan to make this and out of what? I'm not talking about the rails and such but am really asking about the framing and supports. Will you be wielding it together or bolting?

The main frame (grey) is probably going to be welded steel.
The gantry will be 5mm aluminium bolted together
The Y carriage is basically a single peice of 10mm aluminium with bits bolted on.
The Z carriage is a 10mm plat on the bottom and a 5mm on the top.

[FandZ]

I'm on my lunch break so I'll have to keep my post short. I was thinking you'd probably straddle it like you have but I wasn't certain. The length of the Z axis is a little concerning to me. You'll be asking a lof of the round bars to not twist or flex once it is fully extended. Also the forces on your X axes will be great. I don't think, looking forward to you proving me wrong, that you'll be able to mill anything greater than soft woods useing the current setup. I think your Z and possibly your X axis may give you too much sway.

Is there a reason you want so much Z axis travel?

[FandZ]

Also, as far as missing steps for 2 motors on a axis, if you are pushing them so hard for them to miss steps, it just means you are pushing them too hard. Any axis can miss steps if you push it beyond it's spindles cutting ability. I'd give it more thought.

[torleifj]

Stiffness is king!

My bet is that you'll have a hard time cutting aluminum with a 3mm cutter. And PCB cutting will be a pain because motor/spindle vibrations will kill the cutters. With that type of design, I would doubt that 50mm diameter rails would be stiff enough.

My next PCB mill will weigh 100-150kg for that reason. All ways will be supported.

For alu, I use my 1.2 ton Deckel FP2NC. Even that one can bend - it it is fairly stiff though.

/Torleif.

[plasticproducts]

Nice basic design, however, to go past making it work and using it to work, please shorten Z. Anything besides foam will probably twist it off. Bit chatter alone will kill your cut quality.
Why not use track and gears on Y with a drive shaft. Equal torque on all parts when moving on Y.
Keep tunint the design, you're ahead of most I've seen.

[Rarius]

Nice basic design, however, to go past making it work and using it to work, please shorten Z. Anything besides foam will probably twist it off. Bit chatter alone will kill your cut quality.

Done! I have halved the length of the Z axis and doubled up both the ways and the leadscrews. I now have it supported by four 12mm rods and four 12mm leadscrews. Each of the leadscrews passes through a driven nut. One of the nuts is also driven by the stepper.

Why not use track and gears on Y with a drive shaft. Equal torque on all parts when moving on Y.

I don't follow. Are you sure you mean the Y axis.. ie the across axis? I can't see why a track would be any better.

[Rarius]

:stickpoke

Stiffness is king!

My bet is that you'll have a hard time cutting aluminum with a 3mm cutter. And PCB cutting will be a pain because motor/spindle vibrations will kill the cutters. With that type of design, I would doubt that 50mm diameter rails would be stiff enough.

My next PCB mill will weigh 100-150kg for that reason. All ways will be supported.

For alu, I use my 1.2 ton Deckel FP2NC. Even that one can bend - it it is fairly stiff though.

/Torleif.

Thankyou for your input, but not all of us can afford equipment of that scale. :violin:

I am not looking to do aerospace engineering quality work. This is my first CNC machine and I expect it to be a learning experience. I'll be satisfied with 0.1mm accuracy. I don't expect to produce something of commercial grade on my first attempt.

If you are bending a 1.2 ton mill... I suspect you ought to look into how you are using it.. or sounds like "abusing it". :stickpoke

I have shortened the Z-axis (see below) and doubled up the support. Better?

[torleifj]

Thankyou for your input, but not all of us can afford equipment of that scale. :violin:

I have gone all the way from a Sherline with a Dremel to the Deckel. A lot of steps on the way. One of the better steps was buying a like new Denford. The Denfords are very cheap these days and you get a lot of high end hobby parts for the money.

To give you an idea, the Denford weighs something like 250kg, and this was the first one that really did some quality metal work. Still when milling the same jobs on the Deckel, it looks pathetic. Like machining butter.

I also have a PCB mill, that is based on a standard X-Y table. The Z axis was mounted on a 60mm steel pole that was standing on a 40mm alu plate. That was flexing so much that it was useless for alu and it was wasting too many PCB routers. I will redesign it with much larger dimensions soon.

So my advice to you is to build it as stiff as you can possibly afford. Use huge dimensions for the ways and for the structure. Support the corners. Afterall the raw material is the cheap part of a mill and it will pay off.

I am not looking to do aerospace engineering quality work. This is my first CNC machine and I expect it to be a learning experience. I'll be satisfied with 0.1mm accuracy. I don't expect to produce something of commercial grade on my first attempt.

If you are bending a 1.2 ton mill... I suspect you ought to look into how you are using it.. or sounds like "abusing it". :stickpoke

Not at all. All mills are bending, some more than others. The stiffer it is, the larger the chips you cut. The Deckel is very stiff for it's size.

I have shortened the Z-axis (see below) and doubled up the support. Better?

Certainly. Short and large cross section will reward you in the end. Still, I would increase the crossection by a factor of 3.

Good luck with the project. I'm sure that you'll have a lot of fun with it.

/Torleif.

[Rarius]

The Denfords are very cheap these days and you get a lot of high end hobby parts for the money.

I can't quite agree with your definition of "very cheap".

I couldn't find a price in the UK, but I did find one in the US...the cheapest I could find new was $6000! The cheapest I could find on eBay was still £3000 for a second hand micromill.

[torleifj]

Oh, beware! There are many types of Denfords. The micromill is a Sherline, it is not stiff at all. You want a Triac with 30 taper spindle.

They sell for 1500-2000 GBP in UK. But admitted, they are rarely advertized right now.

/Torleif.

[Rarius]

Oh, beware! There are many types of Denfords. The micromill is a Sherline, it is not stiff at all. You want a Triac with 30 taper spindle.

They sell for 1500-2000 GBP in UK. But admitted, they are rarely advertized right now.

/Torleif.

Firstly, £2000 is WAY out of my price range right now. So far I have only spent about £250 and expect to spend about the same again.

Secondly, the point of this project is to build it myself. Anyone can go and buy a ready made machine, but building your own is far more challenging.

[torleifj]

Firstly, £2000 is WAY out of my price range right now. So far I have only spent about £250 and expect to spend about the same again.

Secondly, the point of this project is to build it myself. Anyone can go and buy a ready made machine, but building your own is far more challenging.

Hmm, I understand. I started out the same way. What I realized was that it didn't take long before I got to a point where buying a Denford from the beginning would have been cheaper. Once you go to precision spindles, ballsrews, bearings, motors, motor controls and the like, expenses will go mad.

And buying a mill won't eliminate the need for building your own. I have many machines and I still build more. When you have the first mill, it becomes the tool that makes it easier to build the next.

/Torleif.